Recovering data from corrupted table. Urgent Help!!

ruben20@superguai.com wrote:

Hi:

Is there any way to recover data from a corrupted table? I can only run
SELECTs on certain WHERE conditions.

I cannot vacuum, pg_dump, I've deleted the indexes and try to reindex,
always get error:

ERROR: could not access status of transaction 4244329
DETAIL: could not open file "/usr/local/pgsql/data/pg_clog/0004": No
existe el fichero o el directorio

Thanks a lot.

Again:

create an empty file:

# touch /usr/local/pgsql/data/pg_clog/0004

at this point postgres will complain about the fact that an offset is missing,

at this point fill with 0 your file ( blocks of 8K ) till reach that offset reclaimed.

Regards
Gaetano Mendola

Gaetano Mendola wrote:

ruben20@superguai.com wrote:

Hi:

Is there any way to recover data from a corrupted table? I can only
run SELECTs on certain WHERE conditions.

I cannot vacuum, pg_dump, I've deleted the indexes and try to reindex,
always get error:

ERROR: could not access status of transaction 4244329
DETAIL: could not open file "/usr/local/pgsql/data/pg_clog/0004": No
existe el fichero o el directorio

Thanks a lot.

Again:

create an empty file:

# touch /usr/local/pgsql/data/pg_clog/0004

at this point postgres will complain about the fact that an offset is
missing,

at this point fill with 0 your file ( blocks of 8K ) till reach that
offset reclaimed.

I forgot to suggest you to do:

dd bs=8k count=1 < /dev/zero >> /usr/local/pgsql/data/pg_clog/0004

you have to repeat this command till the offset is covered.

Regards
Gaetano Mendola

Create a file with that name filled with zeros with the same length as
the other files in that directory. That should get you far enough to
dump the data. Then run a complete set of memory and disk checks on
your system...

On Wed, Oct 13, 2004 at 02:56:37PM +0100, ruben20@superguai.com wrote:

Hi:

Is there any way to recover data from a corrupted table? I can only run
SELECTs on certain WHERE conditions.

I cannot vacuum, pg_dump, I've deleted the indexes and try to reindex,
always get error:

ERROR: could not access status of transaction 4244329
DETAIL: could not open file "/usr/local/pgsql/data/pg_clog/0004": No
existe el fichero o el directorio

Thanks a lot.

---------------------------(end of broadcast)---------------------------
TIP 5: Have you checked our extensive FAQ?

http://www.postgresql.org/docs/faqs/FAQ.html

--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/

Hi Gaetano:

This procedure to recover data from a corrupted table should be
documented somewhere... If it is, I could not find it!

Now I wonder if I have lost any data, because after creating the
pg_clog/0004 and running VACCUM everything seems ok.

Thanks a lot for your help.
Ruben.

Thanks Martijn:

It worked. I was having issues with 7.4.2, run memory ans disk checks
and hardware seems ok, that's why I decided to upgrade to 7.4.5, but
unfortunately I'm still experiencing data corruption.

Ruben.

ruben20 wrote:

ruben wrote:

Hi Gaetano:

This procedure to recover data from a corrupted table should be
documented somewhere... If it is, I could not find it!

First of all the table was not corrupted, the glitch was in another
subsystem.
The procedure is documented in the archives :-(

I agree with you but any cure seems worst then the disease. You have understand
why the file 0004 was not anymore there, did you had a power failure for
example ?

Now I wonder if I have lost any data, because after creating the
pg_clog/0004 and running VACCUM everything seems ok.

Normally you didn't lost any data.

Regards
Gaetano Mendola

PS: I had the same error for the first time in my postgres usage life
only after ( some weeks after ) having upgrade from a 7.4.2 -> 7.4.5